---
title: "DiPS Validation"
author: "TJ"
date: "2024-05-29"
output:
  word_document: default
  pdf_document: default
  html_document: default
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)

#install and load packages

install.packages("tidyverse", repos = "https://cloud.r-project.org/")
install.packages("psych", repos = "https://cloud.r-project.org/")
install.packages("GPArotation", repos = "https://cloud.r-project.org/")
install.packages("ltm", repos = "https://cloud.r-project.org/")
install.packages("paran", repos = "https://cloud.r-project.org/")
install.packages("lavaan")
install.packages("lavaanPlot")
install.packages("semPlot")


library(psych)
library(tidyverse)
library(corpcor)
library(GPArotation)
library(ltm)
library(paran)
library(lavaan)
library(lavaanPlot)
library(semPlot)


```

```{r}
#load in data. dat1 = full dataset, dat2 = test-retest participants only

dat <- read_csv("DiPS_data_final.csv")
dat1 <- filter(dat,
               dips_validation_study_survey_1_complete == "2")
dat2 <- filter(dat,
               dips_validation_study_survey_2_complete == "2")
```


```{r}
#demographics

demographics1 <- dat1 %>%
  mutate(ethnicity = dplyr::recode(ethnicity, "1" = "Bangladeshi", "2" = "Chinese", "3" = "Indian", "4" = "Pakistani", "5" = "OtherAsian", "6" = "African", "7" = "Caribbean", "8" = "OtherBlack", "9" = "WhiteAsian", "10" = "WhiteBlackAfrican", "11" = "WhiteBlackCaribbean", "12" = "OtherMixed", "13" = "WhiteEnglish", "14" = "WhiteIrish", "15" = "WhiteGypsy", "16" = "WhiteRoma", "17" = "OtherWhite", "18" = "Arab", "19" = "OtherEthnic")) %>%
  mutate(gender = dplyr::recode(gender, "1" = "Male", "2" = "Female", "3" = "Non-binary", "4" = "Other", "5" = "Prefer_not")) %>%
   mutate(age = dplyr::recode(age, "1" = "18-24", "2" = "25-34", "3" = "35-44", "4" = "45-54", "5" = "55-64", "6" = "65+")) %>% 
   mutate(diagnosis = dplyr::recode(diagnosis, "1" = "Schizophrenia", "2" = "Schizoaffective_disorder", "3" = "Schizotypal_disorder", "4" = "Psychotic_disorder", "5" = "Delusional_disorder", "6" = "Other", "7" = "None")) %>%
  mutate(medication = dplyr::recode(medication, "1" = "Yes", "2" = "No"))


table(demographics1$ethnicity)
table(demographics1$gender)
table(demographics1$diagnosis)
table(demographics1$diagnosis_other)
table(demographics1$medication)
table(demographics1$age)

```

```{r}
#reverse code

dat1 <- dat1 %>%
  mutate(dips6 = dplyr::recode(dips6, "0" = "4", "1" = "3", "2" = "2", "3" = "1", "4" = "0"),
         dips6 = as.numeric(dips6)) %>%
  mutate(dips13 = dplyr::recode(dips13, "0" = "4", "1" = "3", "2" = "2", "3" = "1", "4" = "0"),
         dips13 = as.numeric(dips13)) %>%
    mutate(dips15 = dplyr::recode(dips15, "0" = "4", "1" = "3", "2" = "2", "3" = "1", "4" = "0"),
         dips15 = as.numeric(dips15)) %>%
  mutate(dips22 = dplyr::recode(dips22, "0" = "4", "1" = "3", "2" = "2", "3" = "1", "4" = "0"),
         dips22 = as.numeric(dips22)) %>%
    mutate(scs1 = dplyr::recode(scs1, "1" = "5", "2" = "4", "3" = "3", "4" = "2", "5" = "1"),
         scs1 = as.numeric(scs1)) %>%
  mutate(scs4 = dplyr::recode(scs4, "1" = "5", "2" = "4", "3" = "3", "4" = "2", "5" = "1"),
         scs4 = as.numeric(scs4)) %>%
    mutate(scs8 = dplyr::recode(scs8, "1" = "5", "2" = "4", "3" = "3", "4" = "2", "5" = "1"),
         scs8 = as.numeric(scs8)) %>%
  mutate(scs9 = dplyr::recode(scs9, "1" = "5", "2" = "4", "3" = "3", "4" = "2", "5" = "1"),
         scs9 = as.numeric(scs9)) %>%
      mutate(scs11 = dplyr::recode(scs11, "1" = "5", "2" = "4", "3" = "3", "4" = "2", "5" = "1"),
         scs11 = as.numeric(scs11)) %>%
  mutate(scs12 = dplyr::recode(scs12, "1" = "5", "2" = "4", "3" = "3", "4" = "2", "5" = "1"),
         scs12 = as.numeric(scs12)) %>%
    mutate(ismi2 = dplyr::recode(ismi2, "1" = "4", "2" = "3", "3" = "2", "4" = "1"),
         ismi2 = as.numeric(ismi2)) %>%
  mutate(ismi9 = dplyr::recode(ismi9, "1" = "4", "2" = "3", "3" = "2", "4" = "1"),
         ismi9 = as.numeric(ismi9))

dat2 <- dat2 %>%
  mutate(dips_retest6 = dplyr::recode(dips_retest6, "0" = "4", "1" = "3", "2" = "2", "3" = "1", "4" = "0"),
         dips_retest6 = as.numeric(dips_retest6)) %>%
  mutate(dips_retest13 = dplyr::recode(dips_retest13, "0" = "4", "1" = "3", "2" = "2", "3" = "1", "4" = "0"),
         dips_retest13 = as.numeric(dips_retest13)) %>%
    mutate(dips_retest15 = dplyr::recode(dips_retest15, "0" = "4", "1" = "3", "2" = "2", "3" = "1", "4" = "0"),
         dips_retest15 = as.numeric(dips_retest15)) %>%
  mutate(dips_retest22 = dplyr::recode(dips_retest22, "0" = "4", "1" = "3", "2" = "2", "3" = "1", "4" = "0"),
         dips_retest22 = as.numeric(dips_retest22))
```




```{r}
#create individual measure variables


dips_only <- dplyr::select(dat1,
                study_id, dips1:dips22)

dips_only$dips1sum <- rowSums(dips_only[,-1])


dips13_only <- dplyr::select(dat1,
                             study_id,dips2, dips4, dips6, dips7, dips8, dips9, dips10, dips12, dips13, dips14, dips17, dips21, dips22)

dips13_only$dips13sum <- rowSums(dips13_only[,-1])


dips_retest_only <- dplyr::select(dat2,
                           study_id, dips_retest1:dips_retest22)

dips_retest_only$dips_retestsum <- rowSums(dips_retest_only[,-1])

dips13_retest_only <- dplyr::select(dat2,
                           study_id,dips2, dips4, dips6, dips7, dips8, dips9, dips10, dips12, dips13, dips14, dips17, dips21, dips22)

dips13_retest_only$dips13_retestsum <- rowSums(dips13_retest_only[,-1])

scs_only <- dplyr::select(dat1,
                study_id, scs1:scs12)

scs_only$scssum <- rowSums(scs_only[,-1])


ismi_only <- dplyr::select(dat1,
                study_id, ismi1:ismi9)



ismi_only$ismisum <- rowSums(ismi_only[,-1])


rgpts_only <- dplyr::select(dat1,
                study_id, r_gpts1:r_gpts18)

rgpts_only$rgptssum <- rowSums(rgpts_only[,-1])

hpsvq_only <- dplyr::select(dat1,
                study_id, hpsvq1:hpsvq9)

hpsvq_only$hpsvqsum <- rowSums(hpsvq_only[,-1])


hpsvq_VH <- filter(hpsvq_only,
                           hpsvqsum > 0)

```

```{r}
#split sample

EFA_sample <- dips_only[1:220, ]

CFA_sample <- dips13_only[221:456, ]
  
```


```{r}
#set up factor analysis 

dat1FA <- dplyr::select(EFA_sample,
                 dips1:dips22)

dat_matrix <- cor(dat1FA)

round(dat_matrix, 2)

cortest.bartlett(dat_matrix, n=220)

alpha(dat1FA)

cortest.bartlett(dat1FA)

det(cor(dat1FA))

KMO(dat1FA)
```





```{r}

#factor analysis DiPS13 



dat2FA <- dplyr::select(dat1FA,
                 -dips1,
                 -dips3,
                 -dips5,
                 -dips11,
                 -dips15,
                 -dips16,
                 -dips18,
                 -dips19,
                 -dips20)

cortest.bartlett(dat2FA)

KMO(dat2FA)

det(cor(dat2FA))

cronbach.alpha(dat2FA)


fa1 <- fa(dat2FA, nfactors = 13, rotate = "oblimin")

summary(fa1)

plot(fa1$values, type = "b")

print(fa1$e.values)

```

```{r}
#parallel analysis

paran(dat2FA, cfa = TRUE, graph = TRUE)

```

```{r}
#factor analysis

fa1 <- fa(dat2FA, nfactors = 4, rotate = "oblimin")

fa1$values


factor_loadings <- fa1$loadings
print(factor_loadings)


display_fa <- print.psych(fa1, cut = 0.3, sort = TRUE)

print.psych(fa1, cut = 0.3, sort = TRUE)

```


```{r}
#diagram

colnames(fa1$loadings) <- c("Humanity", "Identity", "Personhood", "Agency")

fa.diagram(fa1, "dat2FA", main = "DiPS Exploratory factor analysis")
```

```{r}
#Confirmatory Factor Analysis



dat1CFA <- dplyr::select(CFA_sample,
                         -study_id,
                         -dips13sum)


model1 <- 'Identity =~ dips2 + dips8 + dips21
Humanity =~ dips7 + dips10 + dips14 + dips17
Personhood =~ dips6 + dips13 + dips22
Agency =~ dips4 + dips9 + dips12'

fit <- cfa(model1, data = dat1CFA)
summary(fit, standardized = TRUE, ci = TRUE, fit.measures = TRUE)

standardizedsolution(fit)


```





```{r}
#CFAdiagram

label_list <- c("dips1", "dips5", "dips12",  
                "dips4", "dips7", "dips10", "dips11",  
                "dips3", "dips9", "dips13",
                "dips2", "dips6", "dips8",
                "Identity", "Humanity", "Personhood", "Agency")

semPaths(fit, 
         whatLabels = "std", 
         fade = FALSE, 
         layout = "spring", 
         nodeLabels = label_list,
         nCharNodes = 10,
         edge.label.cex = 1,
         edge.label.position = 0.5,
         label.cex = 1,
         sizeLat = 8, 
         sizeMan = 6, 
         residuals = FALSE,
         color = list(lat = "skyblue", man = "white"),
         edge.color = "black")
```

```{r}
#floor and ceiling effects

dips_floor <- filter(dips13_only,
                           dips13sum > 0)

dips_ceiling <- filter(dips13_only,
                           dips13sum == 52)
```

```{r}
#density plot

ggplot(dips13_only, aes(x = dips13sum)) +
  geom_density(fill = "skyblue", alpha = 0.5) +
  labs(title = "Density of DiPS scores", x = "Values", y = "Density") +
  theme_minimal()
```



```{r}
#inner join - only use for correlations


validation_sums <- dips13_only %>%
  inner_join(scs_only, by = "study_id") %>%
  inner_join(ismi_only, by = "study_id") %>%
  inner_join(rgpts_only, by = "study_id") %>%
  inner_join(hpsvq_only, by = "study_id") %>%
  dplyr::select(dips13sum,ismisum,rgptssum,scssum,hpsvqsum)

validation_sums_VH <- dips13_only %>%
  inner_join(scs_only, by = "study_id") %>%
  inner_join(ismi_only, by = "study_id") %>%
  inner_join(rgpts_only, by = "study_id") %>%
  inner_join(hpsvq_VH, by = "study_id") %>%
  dplyr::select(dips13sum,ismisum,rgptssum,scssum,hpsvqsum)

```


```{r}
#inner join - RETEST - only for correlations

validation_sums_retest <- dips13_only %>%
  inner_join(scs_only, by = "study_id") %>%
  inner_join(ismi_only, by = "study_id") %>%
  inner_join(rgpts_only, by = "study_id") %>%
  inner_join(hpsvq_only, by = "study_id") %>%
  inner_join(dips13_retest_only, by = "study_id") %>%
  dplyr::select(dips13sum,ismisum,rgptssum,scssum,hpsvqsum, dips13_retestsum)
```


```{r}

#visualisations

ggplot(data = validation_sums, aes(x = dips13sum, y = scssum)) +
  geom_point()+
  geom_smooth(method = lm)

ggplot(data = validation_sums, aes(x = dips13sum, y = rgptssum)) +
  geom_point()+
  geom_smooth(method = lm)

ggplot(data = validation_sums, aes(x = dips13sum, y = scssum)) +
  geom_point()+
  geom_smooth(method = lm)

ggplot(data = validation_sums, aes(x = dips13sum, y = hpsvqsum)) +
  geom_point()+
  geom_smooth(method = lm)

ggplot(data = validation_sums, aes(x = dips13sum, y = ismisum)) +
  geom_point()+
  geom_smooth(method = lm)

ggplot(data = validation_sums_retest, aes(x = dips13sum, y = dips13_retestsum)) +
  geom_point()+
  geom_smooth(method = lm)
```
















```{r}
#reliability

dat_cronbach <- dplyr::select(dips13_only,
                         -study_id,
                         -dips13sum)

alpha(dat_cronbach)

psychosis13<-dat_cronbach[, c(1, 5, 12)]
dehumanisation13<-dat_cronbach[, c(4, 7, 10, 11)]
agency13<-dat_cronbach[, c(2, 6, 8)]
human13<-dat_cronbach[, c(3, 9, 13)]

reliabilty <- psych::alpha(dat_cronbach)
reliabiltypsy<-psych::alpha(psychosis13)
reliabiltyde <- psych::alpha(dehumanisation13)
reliabiltyhu <- psych::alpha(human13)
reliabiltyagen <- psych::alpha(agency13)

```

```{r}
#Cronbach alpha for each scale

Cronbach_voices <- dplyr::select(hpsvq_only,
                 hpsvq1:hpsvq9)
alpha(Cronbach_voices)

Cronbach_paranoia<- dplyr::select(rgpts_only,
                 r_gpts1:r_gpts18)
alpha(Cronbach_paranoia)

Cronbach_compassion<- dplyr::select(scs_only,
                 scs1:scs12)
alpha(Cronbach_compassion)

Cronbach_stigma<- dplyr::select(ismi_only,
                 ismi1:ismi9)
alpha(Cronbach_stigma)
```


```{r}
#correlations


cor.test(validation_sums_VH$dips13sum, 
            validation_sums_VH$hpsvqsum,  
            method = "pearson")

cor.test(validation_sums_VH$scssum, 
            validation_sums_VH$hpsvqsum,  
            method = "pearson")

cor.test(validation_sums_VH$ismisum, 
            validation_sums_VH$hpsvqsum,  
            method = "pearson")

cor.test(validation_sums_VH$rgptssum, 
            validation_sums_VH$hpsvqsum,  
            method = "pearson")

cor.test(validation_sums$dips13sum, 
            validation_sums$rgptssum,  
            method = "pearson")

cor.test(validation_sums$dips13sum, 
            validation_sums$ismisum,  
            method = "pearson")

cor.test(validation_sums$dips13sum, 
            validation_sums$hpsvqsum,  
            method = "spearman")

cor.test(validation_sums$dips13sum, 
            validation_sums$scssum,  
            method = "pearson")

cor.test(validation_sums_retest$dips13sum, 
            validation_sums_retest$dips13_retestsum,  
            method = "pearson")

cor.test(validation_sums$scssum, 
            validation_sums$rgptssum,  
            method = "pearson")

cor.test(validation_sums$scssum, 
            validation_sums$ismisum,  
            method = "pearson")

cor.test(validation_sums$scssum, 
            validation_sums$hpsvqsum,  
            method = "spearman")

cor.test(validation_sums$ismisum, 
            validation_sums$rgptssum,  
            method = "pearson")

cor.test(validation_sums$scssum, 
            validation_sums$ismisum,  
            method = "pearson")

cor.test(validation_sums$ismisum, 
            validation_sums$hpsvqsum,  
            method = "spearman")

cor.test(validation_sums$rgptssum, 
            validation_sums$ismisum,  
            method = "pearson")

cor.test(validation_sums$rgptssum, 
            validation_sums$hpsvqsum,  
            method = "spearman")
```